Ceph » devops

ssh root@node "ceph-disk-prepare --cluster office -- /path/to/mountpoint"

suceeds

Robert Sander wrote:

ssh root@node "ceph-disk-prepare --cluster office -- /path/to/mountpoint"

suceeds

I can confirm that this issue exist on my 3 nodes Ubuntu 12.04 configuration. I have to run the above command on the actual node itself... painful when you have large OSDs. when the ceph-deploy --cluster clustername osd prepare is ran on the admin node the error occurs:

root@cephadmin01:/home/user1/ceph/clusters/openstack# ceph-deploy --cluster openstack osd prepare ceph01:sdb
ceph-disk-prepare -- /dev/sdb returned 1

2013-07-22 16:55:11.107910 7fe3a0a45780 -1 did not load config file, using default settings.
ceph-disk: Error: getting cluster uuid from configuration failed

Alfredo Deza wrote:

Can you confirm the actual output in the node that is failing

I am sorry but we just rebuilt the cluster with the default name 'ceph'.

I am not longer able to reproduce the issue.

Hello! I'm running into similar issue..

***
root@svl-swift-1:/home/ceph# /usr/local/bin/ceph-deploy --cluster alln01ceph -v osd prepare svl-swift-1:sdk --zap-disk
Preparing cluster alln01ceph disks svl-swift-1:/dev/sdk:
Deploying osd to svl-swift-1
Host svl-swift-1 is now ready for osd use.
Preparing host svl-swift-1 disk /dev/sdk journal None activate False
ceph-disk-prepare --zap-disk -- /dev/sdk returned 1
****************************************************************************
Caution: Found protective or hybrid MBR and corrupt GPT. Using GPT, but disk
verification and recovery are STRONGLY recommended.
****************************************************************************
GPT data structures destroyed! You may now partition the disk using fdisk or
other utilities.
The operation has completed successfully.

Caution: invalid backup GPT header, but valid main header; regenerating
backup header from main header.

2013-08-06 00:08:07.118720 7f0d16598780 -1 did not load config file, using default settings.
ceph-disk: Error: getting cluster uuid from configuration failed

ceph-deploy: Failed to create 1 OSDs

***
root@svl-swift-1:/home/ceph# /usr/bin/ceph-conf --cluster alln01ceph osd --lookup fsid
4e79141a-05f2-4e6a-a470-ce54e6236d8e

***
root@svl-swift-1:/home/ceph# cat alln01ceph.conf 
[global]
fsid = 4e79141a-05f2-4e6a-a470-ce54e6236d8e
mon initial members = svl-swift-1, svl-swift-2, svl-swift-3
mon host = 10.29.125.4, 10.29.125.5, 10.29.125.6
auth cluster required = cephx
auth service required = cephx
auth client required = cephx

[client]
rbd cache = true
rbd cache writethrough until flush = true
rbd cache size = 32 MiB
rbd cache max dirty = 24 MiB
rbd cache target dirty = 16 MiB
rbd cache max dirty age = 1

[mon]
osd pool default size = 3

[osd]
osd journal size = 10240
filestore xattr use omap = false
osd mkfs type = xfs
osd mkfs options xfs = "-f -i size=2048 -n size=65536 -l size=65536" 
osd mount options xfs = "rw,noatime,logbsize=262144,logbufs=8" 
filestore flusher = false
osd op threads = 4
osd disk threads = 4
journal aio = true

***

Thanks!
Amit

Amit Vijairania wrote:

Hello! I'm running into similar issue..

***
root@svl-swift-1:/home/ceph# /usr/local/bin/ceph-deploy --cluster alln01ceph \-v osd prepare svl-swift-1:sdk --zap-disk
Preparing cluster alln01ceph disks svl-swift-1:/dev/sdk:
Deploying osd to svl-swift-1
Host svl-swift-1 is now ready for osd use.
Preparing host svl-swift-1 disk /dev/sdk journal None activate False
ceph-disk-prepare --zap-disk -- /dev/sdk returned 1 ********************************************************************
Caution: Found protective or hybrid MBR and corrupt GPT. Using GPT, but disk
verification and recovery are STRONGLY recommended. ********************************************************************
GPT data structures destroyed! You may now partition the disk using fdisk or
other utilities.
The operation has completed successfully.

Caution: invalid backup GPT header, but valid main header; regenerating
backup header from main header.

2013-08-06 00:08:07.118720 7f0d16598780 -1 did not load config file, using default settings.
ceph-disk: Error: getting cluster uuid from configuration failed

ceph-deploy: Failed to create 1 OSDs

***
root@svl-swift-1:/home/ceph# /usr/bin/ceph-conf --cluster alln01ceph osd --lookup fsid
4e79141a-05f2-4e6a-a470-ce54e6236d8e

***
root@svl-swift-1:/home/ceph# cat alln01ceph.conf
[global]
fsid = 4e79141a-05f2-4e6a-a470-ce54e6236d8e
mon initial members = svl-swift-1, svl-swift-2, svl-swift-3
mon host = 10.29.125.4, 10.29.125.5, 10.29.125.6
auth cluster required = cephx
auth service required = cephx
auth client required = cephx

[client]
rbd cache = true
rbd cache writethrough until flush = true
rbd cache size = 32 MiB
rbd cache max dirty = 24 MiB
rbd cache target dirty = 16 MiB
rbd cache max dirty age = 1

[mon]
osd pool default size = 3

[osd]
osd journal size = 10240
filestore xattr use omap = false
osd mkfs type = xfs
osd mkfs options xfs = "-f -i size=2048 -n size=65536 -l size=65536"
osd mount options xfs = "rw,noatime,logbsize=262144,logbufs=8"
filestore flusher = false
osd op threads = 4
osd disk threads = 4
journal aio = true

---

Thanks!
Amit

This should fix it:
--- /usr/share/pyshared/ceph_deploy/osd.py
++ /usr/share/pyshared/ceph_deploy/osd.py
@ -113,6 +113,8 @
args.append('--dmcrypt-key-dir')
args.append(dmcrypt_dir)
args.extend([
'--cluster',
+ cluster,
'--',
disk,
])